318 
ZOOLOGY: KOFOID AND SWEZY 
one of which are attached two of the anterior flagella, while the other 
takes the remaining one and a new anterior flagellum which grows out 
from the blepharoplast. We have not found the blepharoplast to con- 
sist of four basal granules at the roots of the flagella as indicated by- 
Martin and Robertson^ for Trichomonas gallinarum. 
The new chromatic basal rod grows out from the blepharoplast (figs. 
3, 6), apparently independently of the old, as a new structure in the 
membrane below one of the chromatic marginal filaments (fig. 3) and 
distally extends in a new posterior flagellum, thus completing the 
division of the undulating membrane. 
The metaphase is approximated in figure 4 in which, however, the nu- 
clear membrane remains intact, as in fact it does throughout mitosis. 
The two blepharoplasts have migrated to the two poles of the pointed 
ellipsoidal nucleus, and each has divided into a centrosome {cent.), at 
the apex of the spindle, and the adjacent basal granule {has. gr.) to which 
the flagella remain attached. In some instances (fig. 5) this division of 
the blepharoplast is not apparent. No astral rays are evident. Within 
the nuclear membrane faint spindle fibres connect the parting chromo- 
somes to the centrosomes at the poles. Connecting the two blepharo- 
plasts as they migrate to the polar position (figs. 3-6) is a heavy chro- 
matic thread which lies outside of the nuclear membrane. This we name 
the paradesmose, though in origin it may seem to be homologous to the 
central spindle of the metazoan mitotic figure. It is this structure which 
according to DobelP gives rise to the new axostyles of the daughter 
cells. It later disappears (fig. 7) without giving rise to any structures. 
As the split chromosomes swing in to the equatorial plate they assume 
the end-to-end position like that described by Montgomery^ for the 
spermatocytes of Euschistus. Spindle fibres play no part in their split- 
ting. Slight inequality in an "x"-"y" relation between the daughter 
chromosomes is generally evident in the case of the large chromosome. 
There is no evidence, however, that any of the divisions here described are 
maturation divisions, or that their end-to-end position is a telosynapsis. 
The anaphase is accompKshed by the movement of the chromosomes 
to the poles of the spindle into contact with the polar centrosomes and 
by the constriction of the nucleus into a dumbbell shape (figs. 5-6). 
The telophase (fig. 6) results from the final constriction and separation 
of the daughter nuclei which immediately assume a spheroidal form and 
move apart drawing out the paradesmose between their basal granules. 
This is perfectly distinct as an extra-nuclear chromatic thread joining 
the two blepharoplasts or the two basal granules after each blepharo- 
plast parts into centrosome and basal granule (figs. 4-6). The parades- 
